Zalman CNPS7700......... Very disappointed

munichkid · Post by **munichkid** » Wed Jan 19, 2005 4:24 pm

What about using a 2 part epoxy??
We have some stuff at work the is used to hold bolts in plastic on cars.
I have never seen this come off.

Tibors · Post by **Tibors** » Wed Jan 19, 2005 4:30 pm

Using something as permanent as epoxy makes it impossible to ever change the fan again. So unless you can fabricate a new metal bracket, I wouldn't use it.

munichkid · Post by **munichkid** » Wed Jan 19, 2005 4:39 pm

Never though of that. But even if you cannot change the fan, how long does the average overclocker keep the same heatsink? Personally, I am on my 2nd one in a year due to the new stuff that keeps coming out.
By the time you have to chance the fan, a new and improved or better heatsink will be out and we all will be wondering how to "fix" that one up.

Just my 2 cents

Post by **Ralf Hutter** » Thu Jan 20, 2005 6:22 am

munichkid wrote:What about using a 2 part epoxy??
We have some stuff at work the is used to hold bolts in plastic on cars.
I have never seen this come off.

I imagine that would work just fine, although it would couple the fan pretty tightly onto the heatsink, transferring any of the fans' vibrations into the heatsink/mobo. For the ultimate in low noise, I'd prefer to decouple the fan, if possible.

munichkid · Post by **munichkid** » Thu Jan 20, 2005 1:31 pm

This coupled with silicone washers on the 2 screw that hold the chassis to the heatsink should help in eliminating some of the virbations..

On a somewhat related subject, I always wondered if you could use insolation spacers or silicone spacers on the mounting points for the motherboard.

TheWesson · Post by **TheWesson** » Fri Jan 21, 2005 1:23 am

another alternative would be to remove the fan, and have a passive side duct directly above the 7700, and turn all case fans to exhaust, and seal miscellaneous holes.

Upside: you have the combined power of multiple fans pulling air in through the side duct, so this means you can have the other multiple fans at pretty slow speed, so you can have a lot of air with little noise. Also, the Zalman is getting room temp air, which is always nice.

Downside: the airflow is going to be smooth so less efficient in cooling (though I would expect still better cooling per noise than before.) Also, some people like pressure in their case to be positive relative to the environment - this combined with filtered intakes helps keep dust from coming in miscellaneous holes.

anyhow I think this should work well; such a big heatsink should be very good at dumping heat into the incoming airflow, even if the incoming air is a wide flow heading out from the heatsink in all directions. The impedance on the airflow should be pretty small, too, due to the widely spaced fins, so the pressure differential will be small, which will allow even low-pressure (that is, slow and quiet) case fans to work well cooling both your case and CPU.

it sounds like such a good idea I'm sure it's been done a lot before.

lausmeister · Post by **lausmeister** » Sat Jan 22, 2005 7:51 am

I recently bought a 7700 and installed it today. At first I used a very old tube of Arctic Silver 2 thermal compound, but the cooling performance was devastatingly poor to say the least. It couldn't cool my socket 775 P4 3Ghz with fanmate turned to full speed properly when putting the cpu through a 100% load with burnmax. So I thought maybe the Arctic Silver was a little bit to old, so I installed it all again with the thermal grease that came with the 7700. Situation remains the same. 65°C and more under full load with full speed fan. My cpu even starts to emit that high pitch overheat stress signal.
Not even opening the side door seems to help a lot.

So I'd be very interested what you guys think of that.

PS.: On Idle CPU i get 47°C.

EDIT: I just touched it while under full load and the fins are not even really hot. Seems to me the heat is not properly transfered to the heatsink.

Baker · Post by **Baker** » Sat Jan 22, 2005 8:25 am

lausmeister wrote:PS.: On Idle CPU i get 47°C.

EDIT: I just touched it while under full load and the fins are not even really hot. Seems to me the heat is not properly transfered to the heatsink.

Hmm, I would also suspect that it's not seated properly. On my Asus A8V, I had a bit of a fuss removing the built-in mounting bracket on the mobo to replace it with the one that comes with the 7700. I'm not sure what your mounting situation is.

My athlon64 idles at 32C, and doesn't go beyond the mid 40's even when fully stressed for hours. And that's with the fan using 3 or 4 volts and spinning at what I would approximate to be 500 RPM (very very quiet at that level).

Room ambient is ~21C, and the motherboard ambient reads 25C.

I didn't mention it in my previous post, but I did mount mine with Arctic Silver 5. I don't know how much of a real difference that would make, perhaps a few degrees?

Also, I don't know the precise difference in wattage between our processors.

lausmeister · Post by **lausmeister** » Sat Jan 22, 2005 9:21 am

Okay I fiddled around a bit with the 7700 again. I'm absolutly sure it is properly mounted. I'm beginning to think it has something to do with the cleaning product I used to clean the cpu and heatsink. It is an hama optical cleaner and I thought it was just pure isopropyl alcohol, but upon further inspection the bottle states "only use for glass!". So maybe there is something in it to protect the camera lenses it is made for but somehow ruins the cooling performance. I think I'll have to buy something proper and clean the cpu and heatsink thoroughly and install it again.

JanW · Post by **JanW** » Sat Jan 22, 2005 1:28 pm

lausmeister: I was surprised at how sensitive CPU temperature is to what seems like small changes in case airflow. I have a 7000 in a Sonata case (Asus M/B), and the rear exhaust is to the left side of the case, whereas the hot air is generated at the M/B level (to the right side of the case). Even though my situation was quite a bit less dramatic than yours (I only used the fan at the lowest fanmate setting), my temperatures were on the high side compared to those reported by others (63+ deg C on an XP2500+ OC'ed to 185x11, not the system from my sig). Recently I moved the exaust fan from the back to above the M/B (where the PSU would normally go), all the way to the right side of the case, trying to extract the hot air right where the HSF generates it. CPU temperatures went down by 8 deg C, without changing anything on the CPU or its HSF. And this is an undervolted 120mm Nexus as exhaust fan, nothing that moves much air.

The interesting thing is that the case temperatures as reported by the M/B sensor did not change one bit. I guess the sensor is further down in the case, not where the hot air accumulated. So even though it seems most likely that your troubles are related to the HSF, not case airflow, you might want to measure the temperature of the air it takes in. I'd caution against relying on the M/B sensor for case temperature measurements in this context.

lausmeister · Post by **lausmeister** » Sat Jan 22, 2005 1:57 pm

Thanks JanW for the advice I'm definitly going to experiment with airflow while shops are closed until monday when I can get some proper cleaning liquid.

lausmeister · Post by **lausmeister** » Sun Jan 23, 2005 9:25 am

So I took the 7700 off and cleaned it again and installed it several times now. Mounted it in different directions to be sure it sits tight and judging from the dispersion of the thermal grease it definitly does. But Performance is totally unacceptable. I can't run the CPU in full load for more than 5 minutes without it reaching 65°C and thats with the fan to full speed and the caseside open.

I'm really lost here. Is the socket 775 P4 3Ghz really that hot? Or is something wrong. Any suggestion would help.

I'm slowly loosing my faith in that heatsink and already think about returning it, if possible.

Pauli · Post by **Pauli** » Sun Jan 23, 2005 9:39 am

lausmeister- Are you using an Abit motherboard? Some Abits report temps more than 10C too high.

lausmeister · Post by **lausmeister** » Sun Jan 23, 2005 10:36 am

Nope. I've got an Asus P5GD1 and even the Processor itself is emitting an overheating signal. Its a very high piched sound. Otherwise it would at least be possible to ignore the high temperature.

Pauli · Post by **Pauli** » Sun Jan 23, 2005 11:16 am

An ASUS? That's bad -- ASUS boards tend to Under-report temps! There has got to be something wrong with the mounting. I was running a 7000alcu on my P4 Northwood running at 3.28GHz and it never got above 52C under full load. I know the Prescotts run hotter, but they can't be running that much hotter - especially with a 7700.

Tibors · Post by **Tibors** » Sun Jan 23, 2005 11:24 am

If you really want to know how accurate your temps are, then try this:
Calibrate Your CPU Temp Reporting

I'm cooling a socket 478 Prescott 3.0GHz with a Zalman 7000-AlCu and my temps never get that high.

JanW · Post by **JanW** » Sun Jan 23, 2005 11:51 am

lausmeister: Did you run this setup ok with a different HSF before? Just wondering whether there is a way the CPU could have problems. I'm totally shooting in the dark here, as I have never seen a Pentium, but don't they have an integrated heatspreader? Could there be a problem with the contact from that heatspreader to the die? Again, I'm totally clueless and just speculating

If a different HSF worked before, is there a way for you to go back to that?

lausmeister · Post by **lausmeister** » Sun Jan 23, 2005 1:07 pm

Thanks for all the advice I'm continuing my investigations

JanW: I hat the intel stock heatsink working before. So I guess the heatspreader should work correctly. I could go back to that but I'd have to unscrew the motherboard again. I'm trying to avoid that.

My recent guess is, that the 7700 isn't pressed down hard enough. I'm trying to find a way to increase the pressure.

munichkid · Post by **munichkid** » Mon Jan 24, 2005 4:12 am

lausmeister wrote:Thanks for all the advice I'm continuing my investigations

JanW: I hat the intel stock heatsink working before. So I guess the heatspreader should work correctly. I could go back to that but I'd have to unscrew the motherboard again. I'm trying to avoid that.

My recent guess is, that the 7700 isn't pressed down hard enough. I'm trying to find a way to increase the pressure.

The bar where the screws go into can be bend ever so slightly upwards towards the top of the fan vs towards the heatsink side. This will increase pressure. You have to be careful not to bend too much for two reasons. One not being able to screw down the heatsink and MORE importantly, breaking the mb or the cpu.

lausmeister · Post by **lausmeister** » Mon Jan 24, 2005 5:01 am

munichkid: Hm yeah that would work. A friend of mine once crushed his new cpu with a heatsink, so I think I'm going to be extra cautious. To say it with Frys words: "What's the worst thing that can happen?"

thanks

lausmeister · Post by **lausmeister** » Mon Jan 24, 2005 5:51 am

*sigh* increasing pressure doesn't seem to work either. I don't know whats wrong with this thing. Can a heatsink be broken? That's really getting annoying.
I wish I'd have taken the XP-120. If only I could get the 775 conversion plate somewhere here.

Tibors · Post by **Tibors** » Tue Jan 25, 2005 7:55 am

lausmeister wrote:If only I could get the 775 conversion plate somewhere here.

Look here at ichbinleise Austria.

lausmeister · Post by **lausmeister** » Thu Jan 27, 2005 1:51 pm

Hey! Thank you Tibors, thats a wonderful link. Actually that shop seems to be just around the corner for me. It's quiet funny that you from the netherlands have to show me a shop in my neighbourhood.

My only Problem is, that I have the 7700 and don't think that I can return it with the package totally torn apart. I don't want to just blow the 40 Euro I paid for it in the wind and spend another 50 for the xp-120.
I haven't given up on the 7700 yet. Got it to acceptable temperatures, with horrible noise. So I'm in the process of modding it with a papst 4412f/2gl. Experimenting continues. Case airflow definitly is of major importance here. I'll try everthing before I totally give up on the Zalman.

But thanks again anyway. That shop seems to hold a lot of other products I may need, and it's just a few hundred meters away.

Arcanthis · Post by **Arcanthis** » Sat Jan 29, 2005 9:19 pm

This is a real newbie thing, so forgive me if it seems condescending, but you have removed the plastic protector from the bottom of the heatsink (assuming it has one)?

lausmeister · Post by **lausmeister** » Sun Jan 30, 2005 1:42 am

Arcanthis: Hehe it didn't have one to start with, but with the amount of isopropylalcohol I scrubed onto it even plastic would have been rubbed off by now. I appreciate every advice though, don't be afraid to give out newbie advice. The simpler the solution the better

TheWesson · Post by **TheWesson** » Tue Feb 01, 2005 12:00 pm

lausmeister -

Take a look at the base of the HS - is it really flat? Do you feel anything when you scrape a fingernail across it?

Also, if you hold a straightedge (good ruler) against the base and hold it up to the light, do you see a gap? A gap of any size indicates a warped base.

You can smear a little bit oil on the base and put it on a flat sheet of glass (from a picture frame, maybe) and see the contact that it's making from underneath the glass.

Can it be easily rocked a little bit on a flat surface?

Anyhow, the point is, if it's not flat, perhaps some lapping would help.

(Technically, it's not really lapping, just careful sanding of the base, but ...)

lapping:
http://overclockers.com/tips1015/

ps: Also, of course, Prescott CPU's do run very hot and 65C is not surprising (that's a routine temp for a Prescott), but you should be doing better with such a massive heatsink.

ckolivas · Post by **ckolivas** » Tue Feb 01, 2005 12:40 pm

Why not do a simple experiment and check the temperature at load and idle with the side of the case off? That will tell you if you have enough case flow.

lausmeister · Post by **lausmeister** » Wed Feb 02, 2005 8:24 am

TheWesson: Thanks for those tips. I've already tought about a warped base, but your methods to test it are really helpful. I'm definitly checking them out.

ckolivas: There is enough airflow in and out of the case, but a lot of the hot air get's trapped beneath the HS. At the moment I'm planing to build a ducting system to comprehend for that.

Overall I got it working for me. With the fan turning at 1600 i get an idle of 33°C and a load of 55°C. But it's nowhere silent. So I'm going to work on it a little more (e.g. mod it with a 120mm Papst). Maybe a Thermaltight XP-120 would have been less troublesome, but actually it's some kind of challenge for me and starts to make fun.

Zalman CNPS7700......... Very disappointed

CNPS 7700 Performance

Re: CNPS 7700 Performance

Case air flow